The Google zero test: how affiliates rebuild around owned audiences
Search is no longer a business model, only a channel. As AI answers and core updates erode organic traffic, affiliates are buying brands, acquiring communities, renting creators and building products. But as Pierric Blanchet writes, only some of these moves create lasting value from the buy side.
The numbers coming out of digital publishing in 2026 all point in the same way. According to TollBit's State of the Bots data, click-through rates from AI tools to websites that had signed content licensing deals fell from 8.8% in Q1 2025 to 1.33% in Q4 2025. Across AI tools as a whole, the rate dropped from 0.8% to 0.27% between Q2 and Q4. Signing a licence bought publishers visibility, not visits.
France offers an even cleaner experiment. Google rolled out AI Overviews there on 22 July 2026. Two months later, the websites of 40 news brands audited by the ACPM recorded 22.9% fewer visits than a year earlier. Their apps lost only 9.4% on a like-for-like basis. September 2025 was an unusually heavy news month, but August tells the same story: sites fell 18.8%, apps 2.5%.
Same newsrooms, same stories, very different outcomes. The only variable is who owns the relationship with the reader: Google or the publisher.
Some publishers have already drawn the conclusion. GRV Media is the UK sports publisher behind HITC. Its monthly unique users fell by around 65% to 15 million after Google's spam and Discover updates in September 2025, and it cut its headcount from roughly 156 to 94. Its response was Talking Points, a fan-engagement platform that co-founder Vic Daniels described as a "walled garden", built explicitly to prepare for "Google Zero". Daniels has since gone further, setting out his plan on LinkedIn to build a network of independent publishers to fight back against Google.
AI answers replace exactly the informational content that used to feed the funnel, and the commercial results page is increasingly a paid one
Why affiliates are even more exposed
iGaming affiliates sit further down the same funnel, and every major Google update now shows up in their numbers. Gambling.com Group grew 2025 revenue by 30% to $165.4m. That was short of the over 35% it had guided, and the company blamed Google core updates in late 2025 that hurt affiliate rankings.
So far, commercial gambling queries trigger AI Overviews less often than informational searches, and Google keeps paid ads at the top on those terms. That is cold comfort. AI answers replace exactly the informational content that used to feed the funnel ("how does a bet builder work?"), and the commercial results page is increasingly a paid one.
Across the industry, I see four responses.
1. Buying (or building) a brand people come back to
The earliest movers in sports betting understood that the best defence against an algorithm is a reader who types your URL. Better Collective bought:
- HLTV for around €35m in 2020;
- Futbin for around €105m in 2022;
- Playmaker Capital for roughly $250m in 2024, adding more than 200 million monthly visits across the Americas.
North Star Network, the Paris-based group, has built a portfolio of sports media brands across some 30 countries, from Sports Mole in the UK to BasketUSA in France.
The logic is simple. Take an audience that returns every day for scores, transfer news or esports stats, then add betting monetisation on top. Strong odds comparison sites have the same advantage, because checking prices is a habit, not a search.
The biggest expression of this thesis came this year. Genius Sports completed its acquisition of Legend, owner of Covers, Casino.org and Casino Guru, on 1 May 2026. It paid $900m at completion plus an earn-out of up to $300m. The asset was not rankings: Legend reported 320 million visits from 118 million unique visitors in 2025, and more than two-thirds of them returned regularly. Genius presented the deal as a media and attention play, not an affiliate acquisition.
Two caveats. First, a media brand has to stay a media brand. Google's site reputation abuse policy, introduced in 2024, targets third-party commercial content hosted on authoritative domains, so buying a newspaper to bolt casino reviews onto it is not a strategy. Second, audiences notice when editorial turns into a sales funnel. The value lies in the habit, and monetisation that breaks the habit destroys the asset.
The value lies in the habit, and monetisation that breaks the habit destroys the asset
2. Acquiring communities
A second group, mostly mid-sized tier 2 affiliates, is looking at Discord servers, Telegram channels and email databases. On paper, these are owned audiences at a fraction of the price of a media brand. In practice, they are harder to buy than they look.
As with websites and domains, the acquisition itself is often what kills the asset. Moderators and culture hold a community together, not its infrastructure. If it exists only to share bonus codes, members leave as soon as the founder does, or as soon as the new owner pushes offers harder.
Due diligence should cover three things:
- Engagement, not member counts: active members, messages per day, and retention after a change of admin.
- Platform risk: any third-party platform can change its gambling promotion rules overnight.
- Data: in Europe, people give marketing consent to an identified controller, so an email list may need to be re-permissioned after a sale. It can be worth far less after completion than before.
Moderators and culture hold a community together, not its infrastructure. If it exists only to share bonus codes, members leave as soon as the founder does
3. Renting other people's communities
A third approach is to partner with creators: Kick streamers, TikTok accounts, tipsters with loyal followings, and, at the top of the market, star athletes. Better Collective has gone furthest. Its North American talent-led brands, with shows fronted by Shaquille O'Neal, Jalen Brunson and Angel Reese among others, passed one billion views and downloads in July 2025, and the group credits them with a significant share of its 2024 sponsorship revenue of more than €45m. BetMGM, for instance, became presenting sponsor of the Roommates Show with Jalen Brunson and Josh Hart. Better Collective cited "talent-led media" again among the drivers of its return to organic growth in Q1 2026. Note what is being sold here: sponsorship, not only CPA.
For most affiliates, though, creator partnerships replace something that already existed rather than creating an owned asset. They take the place of the old partnerships with mainstream media and the subfolder deals that Google has since penalised. The audience belongs to the creator. At best, the affiliate turns a share of the creator's followers into its own users.
The channel also carries the heaviest regulatory load:
- The Netherlands has banned role-model advertising since 2023.
- X has banned gambling from influencer and paid partnerships.
- Twitch restricted unlicensed slots and roulette streams in 2022, and much of that content moved to Kick.
For a buyer, a creator network is a capability (talent management, compliance, content production), not an asset on the balance sheet, unless the business owns the shows and the long-term talent contracts behind them. It is valued accordingly.
The audience belongs to the creator. At best, the affiliate turns a share of the creator's followers into its own users
4. Building products and new revenue lines
The most ambitious affiliates are building things users come back to by choice.
Products. Score and stats apps are the clearest example. Flashscore ranked first among sports apps worldwide for monthly active users in Sensor Tower's 2026 report. Its owner Livesport has been consolidating the space, adding roughly 25 million monthly users when it bought BeSoccer in 2024. Others are building community products: user reviews, free-to-play prize pools, browser games.
Prediction markets have opened a new front. In September 2026, Legend launched Prediction.com, which gathers pricing from 11 prediction-market apps, including Kalshi and Polymarket. In effect, it is an odds comparison site for event contracts. It comes with a regulatory question mark: in the US, federal and state regulators still dispute sports event contracts, and an affiliate's exposure depends on where its users are.
B2B and new customers. A second route is to sell to someone other than operators. Catena Media's preliminary Q3 2026 revenue fell 29% to €8.3m, as its CEO described a "rapid and fundamental shift" in search and noted that more than 90% of new customers came through CPA deals, which left little recurring revenue to absorb the shock.
Its answer is TraffiqX, a platform it plans to launch in Q1 2027 to handle the whole traffic-exchange chain in iGaming, from selling and setting up traffic deals to measuring, invoicing and reconciling them: a move from lead provider to infrastructure provider.
Studio Leaders, a joint venture launched this week by First Look Games, TAG Media and GameOn, applies the same logic to the supply side. It sells a fully managed affiliate marketing service to game studios, built on First Look Games' tracking of more than 55,000 games across over 1,000 affiliate publishers in 79 markets. The affiliate's know-how becomes the product.
Subscriptions. The evidence here is mixed. Gambling.com Group took subscription revenue from almost nothing to 26% of group revenue in 2025, largely through its acquisitions of OddsJam and OpticOdds. In Q1 2026, its data revenue grew 13% while marketing revenue fell 5%. That is what a hedge looks like.
Oddschecker, by contrast, launched its OC+ subscription in the UK and US in 2024, but made those tools free from 27 October 2026.
My reading: subscriptions work when they sell a tool to a sophisticated bettor or a data feed to a business. They struggle when they put a paywall in front of a funnel whose real revenue comes from CPA and revenue share.
Subscriptions struggle when affiliates put a paywall in front of a funnel whose real revenue comes from CPA and revenue share
What buyers value in 2026 and 2027
Put all of this in front of an acquirer's investment committee, and you can see how far the criteria have moved in the last 18 months.
- Owned acquisition channels. The first question is now the traffic split. What share of sessions comes from direct, app, newsletter or logged-in users, and how did it hold up through the last three core updates? A business that would survive "Google Zero" commands a different multiple from one that would not.
- Community know-how. Buyers pay for teams that know how to run, animate, consolidate and grow a community, because that capability carries across their whole portfolio.
- Owning the audience, not paying for reach. Buyers value revenue bought through paid search, creators or media partnerships as a margin, not as an asset. First-party data and a consented CRM count as assets.
- Capacity to reinvest. Buyers look for businesses that have already paid for their own de-risking, such as an app, a product or a second revenue line, rather than ones that need the buyer's capital to start.
- The CEO's vision. Founders matter more, not less. A CEO who stays on with a credible plan for a riskier environment can carry a transaction. One who still frames the business around rankings can kill it. Buyers will usually ask the founder to stay, and price the deal around that.
- Diversified revenue. B2C and B2B, several operators, several countries. Dependence on one operator or one market remains a classic reason for a discount, all the more when that market's regulation is moving.
- Structure over headline. Buyers are paying less upfront and more over time, through earn-outs and deferred consideration. Even at the top of the market, a quarter of Legend's $1.2bn headline price is contingent. For smaller affiliates, I see earn-outs of 12 to 24 months, tied to revenue, EBITDA or traffic floors, becoming the norm rather than the exception. Sellers should negotiate the metrics, and the control rights that protect them, not just the multiple.
The valuation range makes the point. In the deals I see, affiliate businesses change hands from around 2x EBITDA for search-dependent assets up to 10x for the most premium products. The premium is not for SEO. It is for an audience that comes back on its own.
The Google zero test
Google is not going away, and SEO remains a valuable channel. But on its own, it is no longer a business model. The simplest test for an affiliate preparing for a sale in 2027 is this: if organic search disappeared tomorrow, how much of the audience would still show up? Buyers are already asking that question. The best-prepared sellers will have the answer, backed by data, before it is asked.